Abstract
The immediate priming paradigm can be used to examine how the brain represents scenic layout. A prime picture of a scene is presented for one sec, followed by a brief blank interval, and then a target picture of the scene. Two spatial probes are superimposed on the target and observers make a speeded spatial relations judgment about them. Spatial processing of the target is typically speeded by the prime, relative to controls. We infer that the prime activates a representation relevant to the spatial processing of the target. This representation is broad, fairly detailed, and somewhat high-level, being unaffected by changes in image position and other sensory details (e.g, Sanocki & Epstein, 1997, Psychological Science).
The prime-induced representation could be a scene schema—a network defined by relational information such as spatial relations among objects and surfaces, or by semantic relations such as belongingness. If so, primes should be much less effective when spatial or semantic relations are disrupted.
In the present experiments, we disrupted such relations by cutting the scenic pictures in half and switching the halves. Thus, spatial relations in the middle of scenes were disrupted by interchanging right and left halves (Experiment 1). Semantic relations were disrupted by mixing scene-halves, producing, e.g., a pool/basement scene (Experiment 2). Within each trial, the two sides were constant from prime to target. The spatial relations judgments were local to one or the other half.
Surprisingly, we found that observers were as effectively primed with the mixed stimuli as with normal stimuli. Observers could represent and use primes that were locally organized within halves but inconsistent spatially or semantically between halves. Local spatial organization was indeed critical: When halves switch right and left position between prime and target, the priming effect does not occur (Experiment 3).